Algorithmic inference, political interest, and exposure to news and politics on Facebook
… We ask, how does Facebook algorithmically infer what users are interested in, and how
do interest inferences shape news exposure? We weave together survey data and …
do interest inferences shape news exposure? We weave together survey data and …
Disentangling user interest and conformity for recommendation with causal embedding
… that learns representations where interest and conformity are … users and items with separate
embeddings for interest and … colliding effect of causal inference. Our proposed methodology …
embeddings for interest and … colliding effect of causal inference. Our proposed methodology …
Graph neural news recommendation with long-term and short-term interest modeling
… (2) Our model considers not only the long-term user interest but also the short-term interest.
(3) The topic information incorporated in the heterogeneous graph can help better reflect a …
(3) The topic information incorporated in the heterogeneous graph can help better reflect a …
Controllable multi-interest framework for recommendation
… We propose a greedy inference algorithm to approximately maximize the value function Q(u,…
The items retrieved by different user interests are fed into our aggregation module. After this …
The items retrieved by different user interests are fed into our aggregation module. After this …
Practice on long sequential user behavior modeling for click-through rate prediction
… user Interest Memory Network) to capture user interests from … us to handle the user interest
modeling with unlimited length … realtime inference, with hundreds of millions of users visiting …
modeling with unlimited length … realtime inference, with hundreds of millions of users visiting …
Deep interest evolution network for click-through rate prediction
… to capture the latent user interest behind the user behavior data. … internal cognition, user
interest evolves over time dynamically. … interests to target item and overcomes the inference from …
interest evolves over time dynamically. … interests to target item and overcomes the inference from …
How to make causal inferences using texts
… We next use g to write our causal quantity of interest in terms of the low-dimensional
representation. To make this concrete, consider a case where we have a binary nontext treatment …
representation. To make this concrete, consider a case where we have a binary nontext treatment …
Machine learning at facebook: Understanding inference at the edge
… , leading to overall better quality of user experience. In summary, the significant performance
variability observed for mobile inference introduces varied user experience. If taking a …
variability observed for mobile inference introduces varied user experience. If taking a …
Causal intervention for leveraging popularity bias in recommendation
… popularity bias in the inference stage that generates top𝐾 … a new training and inference
paradigm for recommendation … to discover user real interests and the inference adjustment with …
paradigm for recommendation … to discover user real interests and the inference adjustment with …
Model-agnostic counterfactual reasoning for eliminating popularity bias in recommender system
… To this end, we resort to causal inference which is the science of analyzing the relationship
… though the user is more interested in basketball. Such bias is removed in the inference stage …
… though the user is more interested in basketball. Such bias is removed in the inference stage …